Study sites
The Fram Strait is the only deep connection between the North Atlantic and the Arctic Ocean and plays a significant role in global water mass exchange (Fig. 1). The various topographical structures of the Fram Strait lead to a splitting of the warm and nutrient-rich West Spitsbergen Current, carrying Atlantic water northward, into at least three parts. One part enters the Arctic Ocean north of Svalbard (33%), a second branch flows northward along the north western slope of the Yermark Plateau (45%), and the third part (22%), which for our experiments is the most relevant, recirculates immediately in Fram Strait at about 79° N (Manley 1995, Rudels et al. 2000). This region is characterised by strong annual fluctuations in ice-coverage, although the eastern part of the Fram Strait is generally ice free during the summer months (Rudels et al. 2000).
Experimental set-up
A tripod lander equipped with a baited time-lapse camera (model Simrad Mesotech Photosea 5000), a Photosea 1500SX flash, an acoustic Doppler current profiler (Aanderaa Instruments RCM11), baited traps, glass spheres for buoyancy and ballast weight was used. Two acoustic releases allowed the retrieval of the system after deployment periods lasting between 17.5 and 29.5 hours. The camera view was
centred on bait exposed on a grid which was attached to the lander about 15 cm above the seafloor covering an area of 0.7 m2. Pictures (Kodak Ektachrome 200, 35 mm x 35 mm x 30 m) were taken at 3 minutes intervals. Six experiments were carried out with this configuration during two Arctic expeditions of the German RV "Polarstern" in summer 2000 and 2001 (for details see the respective cruise reports; Krause &
Schauer 2001, Fahrbach 2002). All stations were situated in the Fram Strait, Arctic Ocean (Fig. 1, Table 1).
All traps were made of plastic pipes of 65 cm length with a funnel opening of 15 cm (3.5 cm at the end) and three of them were partitioned into two chambers. While bait was only in one of the chambers, bait odour could penetrate into the neighbouring chamber. About 50 g of fish was used as bait in each trap. Three traps were fixed horizontally below the grid, three others 1.20 m and 1.50 m above the seafloor, respectively.
Figure 1: Locations of stations I to VI in the Fram Strait, Arctic Ocean. Arrows indicate the main bottom current direction during 1997 to 2002 (Schauer et al. submitted, Premke et al. 2003).
2° 4° 6°
78°15' 78°30' 78°45' 79° 0' 79°15'
2000 3000
4000
5000
01 00
I II
III IV
V
VI
Freshly thawed fishes and a crustacean were used as bait at comparable mass but in different composition (Table 1). Aiming to identify food preferences by analyses of the time-lapse photographs we used as standard bait the fish species Salmo trutta (trout) or Solea solea (sole).
Length [mm]Length [mm]
I II III IV V
Station a)
b)
n = 1136
n = 1474
Figure 2: Box and Whisker Plot of lengths [mm] of mature females excluding juveniles (white boxes) and males (grey boxes). a) Eurythenes gryllus, b) Tmetonyx norbiensis .
Additionally, other species were used: Psetta maxima (turbot), Scomber scombrus (mackerel), Pleuronectus platessa (plaice), Molva molva (ling), Belone belone (garfish) and heads of Gadus morhua (cod), and in one case a non-fish-bait, a natant decapod (with a wet weight of ~50 g).
Post processing of photographs
Slides were examined with a stereo microscope (Olympus, 10 x 6.3 magnification) to identify and quantify all visible organisms. Individuals were counted on each slide at the beginning, then on every tenth slide and on every fifth slide during dense population structures, respectively. Amphipods visible on the slides, almost exclusively Eurythenes gryllus were easily identified by characteristics such as body shape, colour (red, white and pink) and eye shape. Photographic identification of individuals was difficult, especially in the case of juvenile Eurythenes gryllus and the adult but smaller sized Tmetonyx norbiensis. This will affect data analysis of photographs. Clearly, the counts of amphipod numbers obtained have to be regarded as a crude minimum estimate. The large number of amphipods attracted in our experiments, as well as their overlapping arrivals and departures after about 4 hours also impeded the determination of the number of frames over which a single individual was present.
The identification was verified by analysis of preserved individuals sampled with the baited traps on the lander. As it was difficult to distinguish between small, presumably juvenile amphipods (< 10 mm) of E. gryllus and amphipods belonging to the genus Tmetonyx norbiensis Oleröd 1987, we counted all amphipods < 10 mm together and used the proportion from captured amphipods of E. gryllus and Tmetonyx norbiensis obtained in each experiment to calculate the respective numbers of these two species.
Abundances in the vicinity of the station were also calculated by applying the relationship n = C tarr-2 (after Priede & Merrett 1996), where C is a constant (for each
experiment) depending on amphipod swimming speed and current speed, and tarr is the arrival time of the first amphipod after the lander reached the sea floor.
Post processing of trap material
All organisms collected with traps were fixed onboard in 4 % buffered formaldehyde. In the laboratory, animals were rinsed in fresh water and identified to species or genus level. Sex was determined by external characters, the length of each individual was measured to the nearest millimetre from the apex of the head to the tip of the telson under a stereo microscope while gently straightening the animal. Final measurements were based on average of three readings. All specimens were blotted dry and their wet weight measured individually on a micro balance. Because traps were lost during experiment VI there was no fixed material available from this experiment.
Specimens with oostegites visible under a stereo microscope were considered to be females. Presence of genital papillae between pereonite 7 and pleonite 1 identified males. Individuals without external sexual characters were considered to be juveniles.
In our study, females and males of Eurythenes gryllus were identified at a minimum length of 16 and 17 mm, respectively. It is likely that some individuals within the juvenile category (16 - 30 mm) were unrecognised females because oostegites are difficult to detect in the very early stage of development. They were included in a second category comprising the female and larger juvenile specimens, assuming that all “juveniles” over 15 mm were unrecognised females.
RESULTS
Lengths and abundances of amphipods
Lander deployments I to V collected 4200 amphipod individuals belonging to three species. About 700 specimens were taken randomly from the total sample for further genetic and lipid analyses, resulting in 3494 amphipods for this study, including 2003 individuals of Eurythenes gryllus, 1483 individuals of Tmetonyx norbiensis and eight individuals of Stegocephalus sp..
Table 2: Mean length, standard deviation and length range of Eurythenes gryllus and Tmetonyx norbiensis, all experiments combined for females, males, juveniles < 16mm and juveniles > 15 mm.
Eurythenes gryllus Length [mm]
Area N Mean S.d. Range
Female Experiment I-V 491 34.51 12.05 16-75
Juvenile > 15mm Experiment I-V 539 19.95 3.48 16-33 Juvenile < 16mm Experiment I, II, IV 328 11.26 2,51 4-15
Male Experiment I-V 645 34.30 9.42 17-56
Tmetonyx norbiensis
Female Experiment I, III, IV 142 22.03 4.93 8-34 Ovigerous female Experiment I, III, IV 929 28.67 3.52 11-38
Juvenile Experiment I 4 10.00 3.94 7-15
Male Experiment I, III, IV 408 24.33 3.65 10-31
Eurythenes gryllus. The mean lengths of males, females, females and juveniles and
juveniles were calculated for all experiments (Table 2). Comparing mean length of adult females (34.1 mm) of all five stations to mean length of adult males (34.3 mm) there were no significant differences (ANOVA, p>0.05). Considering those juveniles larger than 15 mm to be females statistical differences become significant. Among the stations significant differences were found in experiment IV, where females are larger than males and in experiment V, where males are larger than females.
Comparisons were also made for the length range of adult females and males at all of the five locations sampled. Despite similar mean body-length of the two sexes it is
obvious that females grow larger with the largest female collected being 75 mm long and the largest male 56 mm (Fig. 2, Table 2).
The mean body length of all juveniles is about 17 mm and 11 mm without juveniles >
15 mm, respectively (Fig. 3). Juveniles < 16 mm were found in experiments I, II and IV.
No juveniles < 16 mm were found at station II and V (Table 2).
Experiments III and V caught fewer amphipods than I, II and IV. Mean length of females at both stations was similar (34 mm/ 31.6 mm) but the mean length of males (33.8 mm/ 43.2 mm) was significantly different (Fig. 2 and 3). The low number of captured amphipods (of both sexes) was also reflected in the photographic data where also the lowest number of amphipod occurred (Fig. 4).
Tmetonyx norbiensis. The mean sizes of males, females, ovigerous females and
juveniles were determined (Table 2). High numbers of T. norbiensis were captured only in experiments I and IV (Table 3). Samples from experiment III contained only five individuals; three females (two of them ovigerous) and two males, with a mean body length of 26.3 mm (females) and 25.5 mm (males).
The smallest females and males identified were to 8 and 10 mm respectively. Because of a lack of external characters, four individuals were considered as juveniles. Females grow larger than males, with the largest female captured of 38 mm and the largest male of 31 mm (Fig. 5).
0
Figure 3: Length-frequency distributions of Eurythenes gryllus from five stations in Fram Strait based on total body length. Males (black, left figures), females (black, right figures), juveniles
>16 mm (shaded) and juveniles <15 mm (white). Note the different scales of the y-axis.
Spatial distribution and aggregation dynamics of the scavenger community on bait
Scavengers captured with traps or determined on photographs belonged to five taxa (Table 1). The scavenger community on our photographs was dominated numerically by the lysianassoid amphipod Eurythenes gryllus followed by Tmetonyx norbiensis, the latter in considerably lower abundance. The third amphipod species, Stegocephalus sp., appeared rarely (single individuals). Other organisms which were detected on the photographs were zoarcid fish and chaetiliid isopods.
Table 3: Sex ratio of Eurythenes gryllus and Tmetonyx norbiensis for each experiment. N indicates the total number of males and females. The total number of males plus females including juveniles > 16mm and the sex ratio of males and females with juveniles > 15 mm are indicated in brackets.
N* : * N :
Experiment I 191 (430) 1:0.45 (1:2.26) 257 1:1.54
Experiment II 575 (799) 1:0.83 (1:1.54) --
--Experiment III 44 (57) 1:0.46 (1:0.90) 5 1:1.5
Experiment IV 264 (330) 1:0.74 (1:1.18) 1217 1:2.99
Experiment V 54 (59) 1:1.25 (1:1.45) --
--Total 1128 (1675) 1:0.73 (1:1.57) 1474 1:2.63
Eurythenes gryllus Tmetonyx norbiensis
0
* in brackets: including juveniles > 16 mm -- no individuals caught
In five of our experiments the amphipod Eurythenes gryllus was the most abundant species, both in traps and on photographs. In the traps of experiment IV the number of Tmetonyx norbiensis individuals exceeded those of E. gryllus by a factor close to four (Figs. 3 and 5). Although T. norbiensis was captured frequently in two of the six experiments, this species was identified on photographs in much lower abundances.
The zoarcid fish Lycodes cf. frigidus appeared occasionally on photographs in all experiments with a maximum of six individuals per photograph (Table 1). During one experiment two fish entered one of our traps that was oriented towards the camera.
The bait inside the trap was at this point nearly consumed and only a few Eurythenes gryllus were still in the trap (the funnel opening was obviously too large). One fish specimen left the trap after a while but the remaining one was finally (after 8 hours) attacked by the few E. gryllus which were still inside the trap. The subsequent release of body fluids of the wounded fish attracted hundreds of new arriving amphipods and the fish was than skeletonized within 11 hours.
The isopod Saduria sabina and the amphipod Stegocephalus sp. were identified only occasionally on the photographs and only a single individual of Stegocephalus sp. was captured. In all of the experiments amphipods were found on the surface of the bait.
Bait was consumed completely, after a minimum of 12 and a maximum of 28 hours leaving clean skeletons only (Table 1).
First arrivals of single individuals of Eurythenes gryllus were recorded after 15, 12, 24, 21, 21, and 9 minutes, respectively (experiments I - VI, minutes given are tab (time at bottom, Table 1). The maximum number of individuals differed from experiment to experiment (Fig. 4) but we also recorded differences in temporal arrival of individuals which are considered to show two types of feeding aggregations:
Type A: maximum numbers of individuals present between 10 to 24 hours tab with a maximum number of 600 – 800 individuals 0.7 m-2 (Fig. 4 I, II, IV, VI) Type B: maximum numbers of individuals present between 3 to 6 hours tab with a maximum number of 300 – 600 individuals 0.7 m-2 (Fig. 4 III, V)
In both Types the feeding communities were dominated by large individuals during the first hours of tab and the majority of large individuals had left the food fall four hours after bait deployment arrival at the sea floor. A permanent coming and going of individuals suggested a rather short retention period at food falls, but we were unable to identify repeat visits by single specimens.
Generally, large amphipods appeared first followed by a second peak of smaller individuals. The type A aggregation is shown by experiment II in which mainly large individuals dominated the first four hours. About 10 hours later the maximum peak of
amphipods (600 individuals per 0.7 m-2) consisted mainly of smaller individuals (Fig. 4 IV). The current flow during the first four hours (5.4 cm s-1 on average) was only slightly higher than later (4.8 cm s-1 on average). The opposite is illustrated in Fig. 4, experiment III type B aggregation. Current speed was rather constant throughout the experiment (7.6 cm s-1 on average) and the number of amphipods reached a maximum of 300 mostly larger individuals per 0.7 m-2 after 4.75 hours tab. After 6 hour tab the abundance declined to few individuals but than increased again to 150 mostly smaller individuals 0.7 m-2 by 16.5 hours tab. In contrast to experiment II, the majority of small individuals attending experiment III did not account for the maximum abundance during this experiment (Fig. 4 III).
Abundances for Eurythenes gryllus calculated from the time of first arrival were 454 km-2 (station I), 885 km-2 (station II), 134 km-2 (station III), 363 km-2 (station IV), 304 km-2 (station V) and 917 km-2 (station VI), respectively. The stations with high abundances of E. gryllus (station I, II, IV and VI) are Type A aggregations while those with low abundances (III and V) belong to Type B.
Time at bottom [h]
Time at bottom
No. of individuals counted on photographs Current velocity [cm s-1]
Type II
Figure 4: Temporal patterns of numbers of individuals of Eurythenes gryllus (dashed line) as the most abundant scavenger counted on photographs from six stations in Fram Strait and the current velocity (black line) of each experiment (right numbers). Arrows indicate change of current direction.
Food preferences and consumption rate
By counting individuals at bait species over time we identified a pronounced preference for round fish species over flat fish species in Eurythenes gryllus . Among the round species trout will be preferred before ling and other round fish species are attacked with equal facility. In experiment IV the first scavenging amphipods fed on trout, those
following 30 minutes later on ling and about one hour later new individuals arriving at the bait started to feed on sole. Amphipods consumed the trout within 8.5 hours tab, and four hours later the sole was skeletonized. The species which served longest as bait was the ling which needed 22 hours for total consumption (Fig. 6).
The consumption rate differed among stations and ranged from 2500 to 5500 g d-1. Consumption rate at station III was divided into two rates: for the entire experiment including the relatively tiny and light shrimp it amounted to about 2500 g d-1, but to a rate of 5600 g d-1 if only the fish are considered.
Figure 5: Length-frequency distributions of Tmetonyx norbiensis based on total body length for males (black, left figures), females without eggs (black, right figures), ovigerous females (grey, right figures) and juveniles (white, right figures). Note the different scales of the y-axis.
As illustrated in Fig. 4 III and V the number of amphipods increased quickly after deployment of the lander at the seafloor during these two experiments, and rapid consumption of the bait followed so that most of the bait was consumed within 10 hours. This rapid bait consumption at station III and V is related to early peak
abundance of Eurythenes gryllus (Type B). Although in contrast to other deployments the bait wet weight was lower in experiments III and V, i.e. 2 and 2.5 kg, respectively, the consumption rate was high with 5000 g d-1 (experiment V) and even the highest calculated (5600 g d-1 excluding the small shrimp) in experiment III. With bait wet weight ranging between 2.7 to 5 kg in the other experiments, consumption rates in experiments of type A were relatively low (2600 g d-1 up to 3800 g d-1) except station IV (5500 g d-1).
Sex ratio
The ratio of males to females was determined for each location (experiment), and for all stations combined. From a reproductive point of view the most important sexual ratio is the number of mature males to mature females. For Eurythenes gryllus this comparison favours the mature males 1:0.7, all experiments combined (Table 3).
Females of E. gryllus carrying eggs were never captured. The sex ratios of Tmetonyx norbiensis strongly favoured females 2.6:1, all experiments combined (Table 3).
Because of low abundances station III was not considered. Generally, ovigerous females were more frequent than the females without eggs.
There is also a length-sex correlation between the stations, which looks approximately similar. Female body lengths of Eurythenes gryllus of station II, which were relatively small at the 25-75 % level (Box & Whisker Plot) compared to other stations, coincide with small body lengths for males at this station. Both, males and females at station IV contained the largest individuals (Fig. 2). There was, however, a considerable difference between the sexes at station V.
Tmetonyx norbiensis: there is also a length-sex correlation within the stations.
Females and males of T. norbiensis at station I were smaller than those at station IV.
For both species there is a correlation between photographic abundances and
numbers of amphipods captured. Fewer amphipods of Eurythenes gryllus on photographs correspond to fewer amphipods captured in traps (Fig. 4).
Table 4: Grouping of all stations (except station IV) in two different types, based on camera, trap and current meter data analysis.
Type A Type B
High maxima (600 – 800 Ind. 0.7 m-2) Low maxima (300 - 600 Ind. 0.7 m-2)
High calculated abundances (454 – 885 km-2) Low calculated abundances (134 – 304 km-2) Slow increase of scavengers to maximum
(10 – 24 hours)
Fast increase of scavengers to maximum (3 – 6 hours)
Low consumption rate (2600 – 3800 g d-1) High consumption rate (5000 - 5300 g d-1) Inconstant current direction (SSW, SE, NNE) Constant NNW current direction
Juveniles caught No juveniles caught
Tmetonyx norbiensis caught No Tmetonyx norbiensis caught
DISCUSSION
For a long time the seafloor of the deep sea was considered as a monotonous, desert - like environment colonised by only few organisms because of the high hydrostatic pressure, low temperature, absence of light and limited food supply (Svedrup et al.
1942, Dayton & Hessler 1972, Somero et al. 1983). Scientific results obtained during the second half of the last century led to a shift in the understanding of the deep-sea ecosystem (Sanders 1968). The deep-sea benthos was thought to be common and decoupled from processes such as primary and secondary production in the water column. We know today that this is only partly true and that the deep-sea benthos may receive temporarily or spatially relatively larger amounts of organic material from either aggregated phytoplankton (Thiel et al. 1989, Pfannkuche et al. 1999) or carcasses of medium and large sized invertebrates and vertebrates (Smith et al. 1989, Britton &
Morton 1994, Klages et al. 2001) than previously assumed. Studies in the northeastern Atlantic revealed that macroaggregates settling from the euphotic zone at a rate of 100
– 150 m per day lead to a deposit of phytodetritus on the sediment surface. Time-lapse camera experiments in different regions of the world ocean demonstrated that the deep-sea benthic community responds rather quickly to such food supply (Rice et al.
1994). Phytodetrital material is colonized rapidly by Bacteria and Protozoa (flagellates and foraminifers), and is ingested by large deposit feeding animals (Gooday & Turley 1990).
Methodology
Baited time-lapse camera experiments from a useful approach to studying the deep-sea scavenger community because they allow a precise simulation of a naturally occurring event in the deep sea. Most of the studies summarised in Table 1 using baited traps (Thurston et al. 2002) worked on single experiment data sets, with different kind and mass of bait so any differences in composition and succession of the motile scavenger community might have been masked. It should be kept in mind that in our experiments captured organisms permitted reliable species or genus identification.
Natural food falls
The role of carcasses as a mechanism for the transfer of organic matter into the deep sea is still under discussion. Smith et al. (1989, and references cited therein) discussed the migratory routes of grey whales (Eschrichtius robustus) along the west coast of North America and concluded that the random distribution of whale carcasses due to natural mortality would lead to on average distance between falls of about 9 km. In another study Smith (1985) calculated that large nekton falls (2 to 40 kg) contribute only about 4 % of the energy needs of the scavenging ophiuroid Ophiophthalamus
The role of carcasses as a mechanism for the transfer of organic matter into the deep sea is still under discussion. Smith et al. (1989, and references cited therein) discussed the migratory routes of grey whales (Eschrichtius robustus) along the west coast of North America and concluded that the random distribution of whale carcasses due to natural mortality would lead to on average distance between falls of about 9 km. In another study Smith (1985) calculated that large nekton falls (2 to 40 kg) contribute only about 4 % of the energy needs of the scavenging ophiuroid Ophiophthalamus